DESCRIPTION (Applicant's abstract): HIV-associated dementia complex (HADC) is a common disease in the late stages of HIV infection. There are many unanswered questions regarding its pathogenesis, i.e., the relationship of the clinical expression of HADC to the pathological findings of encephalitis and particularly the association of CNS viral load to HADC. Recent studies suggest that single antiretroviral agents (abacavir) can reduce CNS viral load, but it is unknown if reduction of CNS viral load results in the amelioration of dementia. Clinical trials of highly active antiretroviral therapy (HAART) suggest it delays the onset of HADC and can reverse HADC. However, since other studies suggest that there is poor CNS penetration of antiretroviral agents, the improving cognition and attenuating neuropathological features is unclear. The severe combined immunodeficient (SCID) mouse model of HIV encephalitis was developed in our laboratory and recapitulates many of the behavioral and pathological aspects of human HIV encephalitis. Thus, it is a system to study the effects of HAART on abnormal behavior and neuropathology. This model will establish whether HAART reduces viral load and improves the behavioral and pathological features of HIV encephalitis. More importantly, these studies will enable us to begin to elucidate the basic pathogenesis of HADC that may lead to other, more specific treatments. The experimental design incorporates a 2 (HAART vs. vehicle) x 2 (HIV infected vs. uninfected) factorial design to test these two hypotheses: I. AZT + lamivudine + indinavir triple antiretroviral therapy (HAART) will (1) lower the viral load in brain tissue and (2) attenuate the neuropathology common to HIV-infected SCID mice. Specific Aim 1: Determine, after intraperitoneal injections, whether: (A) Mice tolerate HAART at the proposed dosages as measured by general appearance, weight, and activity, (B) Viral titer is reduced in brains of HIV-infected mice treated with HAART versus controls as measured by semiquantitative RT-PCR, (C) Measures of HIV encephalitis (neuropathology-such as numbers of HIV-infected macrophages, gliosis, etc.) are improved in HIV-infected mice treated with HAART versus controls. II. Long-term treatment with HAART will improve the neuropathology and behavioral abnormalities associated with HIV encephalitis. HIV-infected and uninfected SCID mice will be evaluated for: Specific Aim 2: Motor slowing and cognitive dysfunction as indexed by performance during motor activity tests and on a spatial learning/memory task; Specific Aim 3: Viral load determined by semiquantitative PCR for mRNA and immunocytochemistry for infected macrophages; Specific Aim 4: Neuropathology determine by gliosis and neuronal apoptosis; Specific Aim 5: Cytokine activity in brain tissue.